Computing and Informatics (E-Journal - Institute of Informatics, SAS, Bratislava)
Not a member yet
1506 research outputs found
Sort by
Scalable Cloud Application Deployment Service for Versatile Cloud Service Deployment and Configuration
Full text linkWe present a cloud management service called RAIN. It has been designed specifically for versatility and scalability of operation, allowing for the processing of a large number of requests at the same time. Its operation is transactional and controlled by a workflow of operations forming one requisition. Requisitions and their operations can be executed in parallel, allowing for high throughput and scalability of the controlled cloud environment(s). The service is being used in day-to-day operations in a commercial environment. It is also designed for high failure tolerance, which is necessary when operating on third party cloud infrastructures. It has been developed and actively used for several years now, giving us a mature tool with many important features added over time, allowing for practical day-to-day operations. The architecture of the service is open and easily extendable to allow the inclusion of new cloud services of various types -- PaaS providers as well as providers of higher-level services. The service is accessed via an asynchronous REST API. It allows the caller to resume execution and not wait for cloud deployment operations to take an arbitrary amount of time to finish, receiving progress updates via a simple callback REST API
Public Opinion Monitoring and Guidance Analysis in the Process of News Dissemination from the Perspective of Big Data
Full text linkIn order to improve the positive effect of news communication, this paper conducts research on the detection and analysis of public opinion in the process of news communication from the perspective of big data and analyzes the guidance of news communication public opinion. Combined with the actual needs of news dissemination, the numerical accuracy, convergence order, numerical convergence and numerical stability of the SS-CSPH method are mainly analyzed and discussed according to the results of numerical simulation. Moreover, this paper confirms that the smooth function and smooth length do have an impact on the solution of the Strang split-corrected smoothed particle hydrodynamics method (SS-CSPH). In addition, this paper constructs a public opinion monitoring and guidance system in the process of news dissemination. From the simulation evaluation results, it can be seen that the method of public opinion detection and guidance in the process of news dissemination proposed in this paper can play a certain role in the monitoring and guidance of news public opinion
Triple DES with Radial Basis Function Networks for Securing in Process of CCTV Footage Images
Full text linkThe present smart city concept drives academics and urban planners to give residents modern, secure, and sustainable infrastructure as well as a fair level of living. CCTV footage has been employed to improve public safety and wellbeing to fill this demand recognising unusual occurrences. Despite scientific advances, it is challenging and time-consuming to monitor video systems inside of buildings. While cloud computing is unable to support a variety of dispersed IoT setups like wireless sensor networks, local services fall far short of meeting storage requirements. New strategies based on AI and deep learning are gradually replacing conventional computer vision techniques. Modern hardware and software are installed in large data centres, enabling them to process enormous amounts of data. To give cloudlets, fog computing, and multi-access edge computing with intelligent privacy protection for video data while it is being kept, we combine three separate edge computing techniques into a hierarchical edge computing architecture. Then, we offer a straightforward but secure solution. In this study, we look at the creation of a CCTV encryption method that combines RBFN with the Triple Data Encryption Standard (DES) to identify anomalies in intelligent video surveillance. The purpose of this study was to encourage the usage of an algorithm based on RBFN-TDES that provides authenticated encryption rather than the original Data Encryption Standard (DES) method. Because the goal of this investigation was to figure out how the original DES algorithm was cracked, it was unavoidable that hackers would succeed. The suggested technique the security requirements for different encryption properties are also examined and analysed by RBFN-TDES. Finally, the comparison findings between the proposed system and existing systems are presented, along with an examination of each's security characteristics and performance metrics. Our proposed RBFN-TDES model is efficient enough to be used for securing CCTV pictures in a distributed computing environment. To evaluate the effectiveness of our approach, we run large experiments on benchmarks constructed on top of the EPFL dataset. Comparing our strategy to state-of-the-art techniques tested over the datasets, we find a 97.21 % increase in accuracy
Alignment-Based Conformance Checking of Hierarchical Process Models
Full text linkProcess mining has received much attention in the field of business process management. Event logs that are generated from information systems can be correlated with the process models for conformance checking. The process models describe event activities at an abstraction level. However, hierarchical business processes, as a kind of typical complex process scenario, describe sub-processes invocation and multi-instantiation patterns. As existing conformance, checking approaches cannot identify sub-processes within hierarchical process models, they cannot be used for conformance checking of hierarchical process models. To handle this limitation, a definition of hierarchically alignment sequences is presented in this paper. Meanwhile, a novel conformance checking approach for hierarchical process models and event logs is proposed. The proposed method has been implemented within the ProM toolkit, which is an open-source process mining software. To evaluate the effectiveness of the proposed approach, both artificial and real-world event logs are utilized in a comparative analysis against existing state-of-the-art approaches
Parallel Near-Duplicate Document Detection Using General-Purpose GPU
Full text linkIn today's data-rich world, one of the most significant challenges is efficiently identifying near-duplicate data, especially when integrating data from various sources. Identifying near-duplicate documents applies to any content and has been widely used to enhance the efficiency of search engines, identify plagiarism or spam, and so on. Even smaller or specialized search engines can benefit from knowledge about near-duplicate documents. Shingling and MinHash are two state-of-the-art approaches to detecting near-duplicate documents. However, there are not many attempts to improve the performance of this locality-sensitive hashing technique. In this research paper, we propose a parallel implementation of the MinHash algorithm for near-duplicate document detection utilizing the immense parallelism offered by general-purpose GPUs. Experimental results show that the GPU-based parallel solution is far more cost-effective than the CPU-based sequential and parallel solutions
Autoscaling Method for Docker Swarm Towards Bursty Workload
Full text linkThe autoscaling mechanism of cloud computing can automatically adjust computing resources according to user needs, improve quality of service (QoS) and avoid over-provision. However, the traditional autoscaling methods suffer from oscillation and degradation of QoS when dealing with burstiness. Therefore, the autoscaling algorithm should consider the effect of bursty workloads. In this paper, we propose a novel AmRP (an autoscaling method that combines reactive and proactive mechanisms) that uses proactive scaling to launch some containers in advance, and then the reactive module performs vertical scaling based on existing containers to increase resources rapidly. Our method also integrates burst detection to alleviate the oscillation of the scaling algorithm and improve the QoS. Finally, we evaluated our approach with state-of-the-art baseline scaling methods under different workloads in a Docker Swarm cluster. Compared with the baseline methods, the experimental results show that AmRP has fewer SLA violations when dealing with bursty workloads, and its resource cost is also lower
Intelligent Cockpit Application Based on Artificial Intelligence Voice Interaction System
Full text linkWith the rapid development of computer technology, intelligent cockpit has increasingly become a trend in on-board related equipment. The intelligent cockpit system can not only provide auxiliary support for users in the driving environment, but also greatly improve driving efficiency. With the advancement of technology and the improvement of people's material lives, people's demand for driving environment safety and intelligence is gradually increasing. In order to effectively enhance the interactivity of the intelligent cockpit and ensure the safety of users' driving, this article effectively analyzed the performance and functional requirements of the intelligent cockpit, and combined speech recognition algorithms to construct an artificial intelligence (AI) speech interaction system. On this basis, this paper provides an in-depth study of intelligent cockpit design. To verify the effectiveness of the intelligent cockpit application, it was tested from four aspects: system operation accuracy, command response time, system security, and user experience. The test findings showed that at the level of system command response time, the response time of driving assistance commands, entertainment media commands, and traffic information commands was basically concentrated below 300 milliseconds. From the specific test results, it can be seen that the intelligent cockpit based on the artificial intelligence voice interaction system can quickly respond to users' different command needs through precise speech recognition
BDANet: Boosted Dense Attention Hierarchical Network for Image Denoising
Full text linkDeep convolutional networks have been widely applied in image denoising tasks with great success. However, many denoising models extract more feature information by increasing the network depth, which does not fully utilize the shallow features, but also makes it difficult to obtain accurate noise information. In this paper, we introduce a novel modified U-Net structure-based boosted dense attention neural network (BDANet) specifically designed for image denoising. The convolutional block within the encoding layer of BDANet incorporates dense connections and residuals, effectively circumventing the vanishing gradient issue through feature reuse and local residual learning. A boosting strategy is employed in the decoding layer to augment residual information in the noise map. To adeptly process edge details in images, BDANet deploys a polarized self-attentive mechanism to direct the densely connected blocks for depth feature extraction. The network is trained with Gaussian noise at random noise levels in the study to make it flexible to handle images with a wide range of noise levels. In experimental comparisons involving additive Gaussian noise, BDANet outperformed conventional denoising networks and attained competitive results relative to state-of-the-art image denoising networks, with an average improvement of approximately 1.03 dB in terms of PSNR values. Visualization results show that the image after denoising by BDANet network is sharper and richer in texture details than other methods
Effective Lightweight Dual-Path Shift Compensation Network for Image Super-Resolution
Full text linkIn this paper, we propose a lightweight dual-path convolutional neural network for image super-resolution (SR). We introduce shift convolution and propose a shift-channel attention (shift-ca) mechanism to build an effective network. Shift-ca produces an attentional map with a larger field of view, and its formulation is similar to channel attention and spatial attention. In addition, we propose the Local Shift-Channel Attention Feature Extraction (LCFE) module as the main part of the Dual Path Shift Attention Block (DPSAB). Using the dual-path structure allows us to reduce the network depth and retain more original features for the subsequent up-sampling compensation operation. In the final HR reconstruction module, we combine the nearest neighbor upsampling layer, convolutional layer, and activation layer to form the compensated nearest neighbor upsampling module (C-NUM) to improve the reconstruction quality with a small parameter cost. Our final model is the Dual Path Shift Attention Network (DPSAN), and it achieves similar performance to the lightweight network WMRN (36.38 % for WMRN) with only 195 k parameters. Applying our module to the EDSR-baseline also yielded good results. The effectiveness of each proposed component was verified by an ablation study.
HRN: Haze-Relevant Network Using Multi-Object Constraints for Single Image Dehazing
Full text linkIn recent years, some deep learning dehazing methods based on atmospheric scattering model mostly solve the dehazing results by using depth convolution neural networks (CNNs) to estimate the medium transmission map in the model. However, these methods usually ignored the potential correlation between the transmission map and the atmospheric light in the atmospheric scattering model, which can lead to colour distortion and incomplete dehazing in the dehazing results. To address this problem, this paper first presents a novel Haze-Veil model to increase the correlation between the model parameters by constructing an atmospheric veil term. Then, based on the proposed model, a haze-relevant end-to-end network (HRN) is designed to estimate the parameters of this model and directly output the final clear image. In addition, a cost function is designed by defining multi-object constraint cost functions to further establish the connections between the statistical attributes of the hazy image and the out of HRN. Experiments on benchmark images, which include synthesized and real images, show that HRN effectively removes haze and outperforms most of the existing and state-of-the-art dehazing methods